LinkedEarth / PaleoBooks

MIT License
2 stars 0 forks source link

Replication study inspired by Steinmann et al 2022 #12

Open CommonClimate opened 4 months ago

CommonClimate commented 4 months ago
  1. Goal: Hydroclimate PaMoDaCo over the neotropics Last Millennium
  2. Description: Use the PRECIPd18O data from Steinmann et al 2022, load into pyleo EnsembleSeries, run MC-PCA and compare to iCESM last millennium simulation (available here and maybe through globus). The idea would be to showcase how to access paleo obs and model data and run a fairly complex (age-aware) analysis and reproduce something that looks like one of the figures in the paper.
  3. References/data sources: Steinmann et al 2022, iCESM
  4. Keywords (science): hydroclimate, MC-PCA
  5. Keywords (tech): pangeo-forge, ARCO data, globus
CommonClimate commented 4 months ago

Data can be found here. @fzhu2e do you know how to turn those fields to d18O? How did you do that for the data you archived as part of cfr?

fzhu2e commented 4 months ago

Data can be found here. @fzhu2e do you know how to turn those fields to d18O? How did you do that for the data you archived as part of cfr?

Hi Julien, here's the recipe:

        p16O = ds['PRECRC_H216Or'] + ds['PRECSC_H216Os'] + ds['PRECRL_H216OR'] + ds['PRECSL_H216OS']
        p18O = ds['PRECRC_H218Or'] + ds['PRECSC_H218Os'] + ds['PRECRL_H218OR'] + ds['PRECSL_H218OS']

        p16O = p16O.where(p16O > 1e-18, 1e-18)
        p18O = p18O.where(p18O > 1e-18, 1e-18)

        d18Op = (p18O / p16O - 1)*1000
CommonClimate commented 4 months ago

Most excellent! Thank you so much for the trick. Immensely helpful!

khider commented 3 months ago

One problem with this: data size. Each text file is around 200-500MB because of the age models. I propose we use the median age model for the cookbook so we can get a smaller file.

We may need to have a "smaller" example (i.e., less proxy) to do MC-PCA. Thoughts?

CommonClimate commented 3 months ago

yes, we can streamline on a subset of proxies, or maybe a subset of the age ensembles?

khider commented 3 months ago

I'm going to do median for the Pythia Cookbook and we can figure out how to store the data in LiPD and on the graphDB later on.

khider commented 3 months ago

@fzhu2e Trying to figure out the numbering system at NCAR?

image

Which one corresponds to the all forcing?

khider commented 2 weeks ago

Done for Project Pythia. Needs to be included in our gallery

https://github.com/ProjectPythia/paleoPCA-Cookbook